I was not sure how to present this issue so I was also not sure how to search for it on the forum.
Three times now we have seen thin volumes suddenly expand in vCenter to the maximum size they were created as. Kind of like they suddenly went from Thin to Thick volumes.
There was no additional usage on the actual OS drive to account for this sudden growth. So far this has only happened with Windows virtual machines. This has happened on three different machines.
For Example (with the latest event).
We had a 6TB Thin volume that had just been created and had a few folders on it amounting to about 800MB. Not long after this volume was created it suddenly started reporting as the full 6TB in vCenter and the host. Checking the metadata on the VMDK confirmed it was a thin drive and checking the OS still showed only the original 800MB of data. Drive was not slated to go into production until later.
The first time this happened I opened a support ticket. After some research and poking, support believed it was a block count issue between the OS and vCenter. They suggested a Windows OS defrag of every affected drive, then a re-check/refresh of the Datastore Capacity to see if that started to get them back inline. If that did not resolve it they had me run the exscli umap command on each of the affected datastores (on the host that was running the VM at the time). This resolved the issue the first and the second time this happened.
The third time this happened we ended up having to destroy the volumes and recreate/restore them.
Has anyone else had to deal with this before?
If so what has your fix been?
Is there any information on what might trigger this?
Thanks much!
William
Where you saving logs
Are you talking about VMWare logs or logs on the OS? Not sure I follow.
William
Welcome to the Community,
growing of a thin provisioned virtual disk is usually caused by some guest OS (guest file system) actions. However, I can't unfortunately tell you what exactly is going on.
Some question though if you don't mind:
André
Not sure, but it may be worth reading https://kb.vmware.com/s/article/56608 to see whether this is what causes the issue in this case.
André
VMware logs